Agent Based Modelling in Biology

Background

In Biology agent based modelling has traditionally been used at a high level, modelling bird flocking or social insect behaviour. With increased computer power, the advancement of molecular biology and the growing interest in systems biology, agent based modelling is now being commonly applied at the molecular and cellular level, with huge implications for the medical world.

The cell is the fundamental unit of Biology. DNA encodes all the information necessary for life, but the cell is (apart from viruses) the smallest entity with with this information can be implemented. It is through the cells processing of DNA that all subsequent properties of life emerge. This is just one field in which agent based modelling can be such a powerful tool in Biology. It lets us model each cell individually and view for ourselves the results of interactions, without us implicitly defining them beforehand.

All characteristics of multicellular lifeforms are emergent properties of the interactions of their building blocks and the environment. There is no governing voice from above directing the cells how to arrange themselves in order to create the shape of a dog, for example. The dog 'emerges' as the cells replicate and interact with each other and their physical environment. Of course their are very strict rules governing the interactions of the cells, essentially being just the physical and chemical limitations of the cells in their current state, and appearing as the direct result of natural selection working over time. But nowhere in the DNA is the 'image' of the dog visible itself, it is an emergent property of the mass of cells, all with identical information but each continuously responding to limitations caused directly or indirectly by their very neighbours.

Agent based modelling works in exactly the same, and enables us to observe emergence in the same way it happens in life. For example, using attainable information such as certain properties of a cell type, we can then predict things such as the outcome over time of tissue growth, this having a direct relevance to cancer formation. Emergence can produce extremely complicated patterns and interactions which are often impossible to predict in a bottom up approach, and hard to explain from a top down approach. Agent based modelling enables us to simulate biological phenomena from a bottom up perspective,. By altering the model and comparing results to real experimental data, we can discover specific mechanisms of development which cannot be observed in vivo.

Click here for a list of papers relating to this growing application of agent based models in molecular and cellular biology.